Model Insights Blog

Baseline Model

Logistic Regression

A Simple but Strong Start

Despite its simplicity, Logistic Regression gives fast, interpretable predictions. It shows which words or features influence the result — making it the perfect baseline for understanding what works before moving to more complex models.

⚡ Fast inference 🔍 Interpretable 📊 87% accuracy

Ensemble

Random Forest

The Ensemble Powerhouse

A collection of decision trees working together. Each tree votes, and the forest averages the results — reducing overfitting and capturing subtle signals from word frequency and structure to uncover hidden misinformation.

🌲 Ensemble method 🛡️ Robust 📊 89% accuracy

Deep Learning

LSTM

Deep Context Understanding

Long Short-Term Memory networks understand sequences — looking at words in context, not just in isolation. LSTM detects tone, intent, and sentence structure, giving a deeper view of a claim's truthfulness for complex sentences.

🧠 Neural network 📝 Context-aware 📊 91% accuracy

Probabilistic

Naive Bayes

Fast & Probabilistic Classifier

Calculates the probability that a statement is true based on the presence of certain words. Incredibly fast and surprisingly accurate on clean datasets — perfect as a lightweight screener that quickly flags potentially false claims.

🚀 Ultra-fast 📐 Probabilistic 💡 Explainable